1

AEM01423-13 Revised 1

2

Real-time PCR for quantitative analysis of human commensal Escherichia coli 3

populations reveals a high frequency of sub-dominant phylogroups 4

5

6

Mounira Smati1,2,3, Olivier Clermont1,4, Frédéric Le Gal3, Olivier Schichmanoff3 7

Françoise Jauréguy1,2,3, Alain Eddi5, Erick Denamur1,4, and Bertrand Picard1,2,3 for the 8

COLIVILLE group 9 1UMR-S 722, INSERM, Paris, France 10 2UMR-S 722, Université Paris Nord, PRES Sorbonne Paris Cité, Bobigny, France 11 3APHP, Hôpitaux Universitaires Paris Seine Saint-Denis, Site Avicenne, Bobigny, France 12 4UMR-S 722, Université Paris Diderot, PRES Sorbonne Paris Cité, Paris, France 13 5Département de Médecine Générale, Université Paris Diderot, PRES Sorbonne Paris Cité, 14

Faculté de Médecine, Paris, France 15

16

17

The COLIVILLE group is composed of the following medical practitioners who were 18

involved in the recruitment of subjects for the study: Monique Allouche, Jean-Pierre Aubert, 19

Isabelle Aubin, Ghislaine Audran, Dan Baruch, Philippe Birembaux, Max Budowski, Emilie 20

Chemla, Alain Eddi, Marc Frarier, Eric Galam, Julien Gelly, Serge Joly, Jean-François Millet, 21

Michel Nougairede, Nadja Pillon, Guy Septavaux, Catherine Szwebel, Philippe Vellard, 22

Raymond Wakim, Xavier Watelet and Philippe Zerr. 23

24

Corresponding author: Erick Denamur, UMR-S 722, INSERM, Université Paris Diderot, Site 25

Xavier Bichat, 16, rue Henri Huchard, 75018 Paris, France. Tel: 33 1 57 27 77 39; email: 26

erick.denamur@inserm.fr 27

28

Key words: Escherichia coli, B2 phylogroup, microbiota, quantitative PCR, commensal 29

30

Running title: A quantitative study of human commensal Escherichia coli 31

32

33

Copyright © 2013, American Society for Microbiology. All Rights Reserved.Appl. Environ. Microbiol. doi:10.1128/AEM.01423-13 AEM Accepts, published online ahead of print on 14 June 2013

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

2

ABSTRACT 34

35

Escherichia coli is divided into four main phylogenetic groups which each exhibit 36

ecological specialization. To understand the population structure of E. coli in its 37

primary habitat, we directly assessed the relative proportions of these phylogroups from 38

the stools of 100 healthy human subjects using a new real-time PCR method, which 39

allows a large number of samples to be studied. The detection threshold for our 40

technique was 0.1% of the E. coli population, i.e. 105 CFU/g of feces; in other methods 41

based on individual colony analysis the threshold is 10%. A single, two, three or four 42

phylogenetic groups were simultaneously found in 21%, 48%, 21% and 8% of the 43

subjects, respectively. Phylogroups present at a threshold of less than 10% of the 44

population were found in 40% of subjects, revealing high within-individual diversity. 45

Phylogroups A and B2 were detected in 74% and 70% of the subjects, respectively; 46

phylogroups B1 and D were detected in 36% and 32%, respectively. When the 47

phylogroup B2 was dominant, it tended to not co-occur with other phylogroups. In 48

contrast, other phylogroups were present when phylogroup A was dominant. These data 49

indicate a complex pattern of interactions between the members of a single species 50

within the human gut, and identify a reservoir of clones which are present at a low 51

frequency. The presence of these minor clones could explain the fluctuation in the 52

composition of E. coli microbiota within single individuals that may be seen over time. 53

They could also constitute reservoirs of virulent and/or resistant strains. 54

55

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

3

INTRODUCTION 56

57

Escherichia coli is the most common commensal aerobic bacteria in the human gut 58

microbiota (1); it is also the Gram negative bacillus most frequently implicated in human 59

extraintestinal infections (2). This apparent paradox, which characterizes opportunistic 60

pathogens, is classically associated with host deficiencies (for example, as a result of immune 61

compromising illness, surgery or catheterism). However, there are still questions concerning 62

the role of the intrinsic properties of the parasite and the conditions which may cause 63

commensal intestinal E. coli strains to become extraintestinal pathogens. 64

Four main phylogenetic groups named A, B1, B2 and D have been distinguished 65

among E. coli species (3, 4) and distribution within these groups appears to correlate with the 66

origin of the strains. Phylogroup B2 has been shown to include extraintestinal virulent strains 67

(ExPEC) which express numerous virulence factors (5, 6) whereas the phylogroups A and B1 68

contain mostly human and animal commensal strains, respectively (7, 8, 9). However, more 69

recently, an increase of B2 phylogroup strains was observed in human commensal strains 70

originating from industrialized countries (10, 11, 12, 13). 71

Little is known about the diversity, transmission and persistence of E. coli commensal 72

strains within human populations (1). An individual can be colonized by more than one 73

distinct strain at any given time (14, 15, 16). However, the few available studies that describe 74

strain diversity are based on the arbitrary selection of colonies from a plate inoculated with 75

fecal samples. The typing methods used included O-typing, biotyping (14, 17), multi-locus 76

enzyme electrophoresis (18, 19, 20), and genotyping methods such as pulse-field gel 77

electrophoresis (21), ERIC-PCR (22, 23), random-amplified polymorphic DNA (RAPD) 78

alone (24) or combined to biochemical fingerprinting (25), ribotyping (26) and triplex 79

phylogrouping PCR (27). The studies included a small number of subjects and/or selected a 80

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

4

relatively small number of colonies on which to base their investigation of E. coli population 81

structure. The ability to accurately characterize E. coli strain diversity in human subjects, and 82

the likelihood of identifying a particular E. coli strain, are directly related to the number of 83

colonies sampled and the underlying prevalence of the strain (21, 26). For example, for a 90% 84

likelihood of identifying a strain present in ≥10% of colonies, 22 colonies must be sampled 85

(21). New approaches are therefore needed to improve the sensitivity of detection. The 86

detection of fecal strains present at low frequencies is necessary to understand within-87

individual fluctuations in strain composition which can be observed over time (18, 20, 28, 88

29). 89

To get insights in the population structure of E. coli in its primary habitat, we assessed 90

the relative proportions of the main E. coli phylogenetic groups in the stools of 100 healthy 91

human subjects; their intestinal microbiota had undergone minimal pathological and medical 92

perturbations. A new, rapid and sensitive real-time PCR strategy, using the same targets as 93

triplex phylogrouping PCR (4), was applied on samples from the subjects. 94

95

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

5

MATERIALS AND METHODS 96

97

Subjects. From May 2009 to December 2011, we recruited 100 healthy human subjects from 98

the region of Ile-de-France (Paris and its suburb area, France). All participants lived in the 99

community and volunteered to self-collect a fecal swab sample. The subjects had no history 100

of gastro-intestinal disease, no symptoms of immunodepression, had not received antibiotic 101

therapy in the previous month and had not been hospitalized in the three months preceding 102

inclusion. Written informed consent was obtained from each participant and the study was 103

approved by the local ethics committee (CCTIRS n° 09.243, CNIL n° 909277, and CQI n° 104

01-014). 105

106

Samples and characterization of isolated strains. Fecal samples were self-collected by the 107

subjects using Amies transport medium swabs (Venturi Transystem, Copan, Brescia, Italy) 108

and sent by mail to the Avicenne hospital laboratory. The swabs were discharged in glycerol 109

stock solution (Cryobank, Biovalley, Marne la Vallée, France) and stored at – 80°C until use. 110

The quantity of feces in 0.2 ml of solution was estimated in preliminary experiments to be 10 111

mg: E. coli counts from 10 swabs was compared to that from corresponding fresh feces which 112

were chosen for a variety of consistency (data not shown). The stool-containing suspensions 113

were plated onto Drigalski agar plates. After 24 h of incubation at 37°C, one colony was 114

randomly picked, identified as E. coli using API20E (BioMérieux, Marcy l’Etoile, France) 115

and stored in glycerol stock solution. Twenty to 30 E. coli colonies were randomly picked 116

from each of 20 original plates and stored in glycerol stock solution at – 80°C to serve as 117

controls. The phylogenetic groups (A, B1, B2 and D) of the randomly picked E. coli strains 118

were determined by the triplex PCR method (4) and by the recently reported quadriplex 119

method which also detects Escherichia clades strains (13, 30). Strains were tested for their 120

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

6

antibiotic susceptibility using the disk diffusion method according to the 2012 121

recommendations of the “Comité de l'Antibiogramme de la Société Française de 122

Microbiologie” (Antimicrobial Committee of the French Society for Microbiology) 123

(www.sfm-microbiologie.org). The following antimicrobial agents were tested: amoxicillin, 124

amoxicillin-clavulanic acid, cefoxitin, ceftriaxone, amikacin, ofloxacin, and 125

trimethoprim/sulfamethoxazole. 126

127

DNA extraction from feces. The QIAamp DNA Stool Minikit (Qiagen, Courtabœuf, France) 128

was used to extract DNA from 0.2 ml of frozen stool sample according to the manufacturer’s 129

recommendations with modifications. An internal DNA control corresponding to the R’ 1 130

region of the hepatitis delta virus (HDV) genome prepared as previously described (31) was 131

added, before extraction, at the quantity of 105 copies per sample. The DNA was eluted in a 132

final volume of 200 µl and stored at – 80°C. 133

134

Quantitative PCR assay. 135

(i) Design of primers and phylogroup-specific probes. The sequences used in the 136

triplex PCR phylogrouping method (4) (TspE4.C2, chuA and yjaA) were obtained from 137

GenBank, EMBL and the Broad-Institute (www.broadinstitute.org) databases (from the whole 138

genome sequences of Escherichia strains) and aligned using the Clustal W program (provided 139

by the European Bioinformatics Institute). The strains were assigned in vitro or in silico to 140

their corresponding phylogroup by the triplex PCR method and/or by multi-locus sequence 141

typing (MLST). Primers were designed to anneal to conserved sequences, whereas the probes 142

were designed to target unique sites allowing specific detection of each phylogroup with the 143

Primer Express 3.0 software (Applied Biosystems, Villebon Sur Yvette, France) and the 144

Primer3 Plus online interface (32) (provided by the Whitehead Institute). The list of strains 145

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

7

tested in silico is given in Table S1. Two strains belonging to phylogroup D (TA255 and 146

TA280) have a phylogroup F chuA sequence (probably resulting from a horizontal gene 147

transfer) and were identified as B2 by our method. Primers and probes were manufactured by 148

Sigma-Aldrich (Lyon, France) and Applied Biosystems, respectively. All primers and probes 149

used are listed in Table 1. 150

151

(ii) Phylogroup-specific real-time PCR. Purified DNA (20 µl) was added to 30 µl of 152

PCR mixture containing 25 µl of TaqMan Universal PCR master mix II (2X) (Applied 153

Biosystems), 300 nM of each primer, 100 nM of fluorescent probe and bovine serum albumin 154

at the final concentration of 0.1 µg/µl (New England Biolabs, Evry, France). However, for the 155

B1 phylogenetic group, primers (TspFW and TspRV) and probe (pTspE4.C2B1) were used at 156

900 nM and 250 nM, respectively. The reaction mixture was heated to 50 °C for 2 min 157

(initiation step) and then 95 °C for 10 min. This was followed by 45 cycles of amplification 158

which each consisted of 15 s at 95 °C and 1 min at 60 °C. A no-template control and a 159

positive control sample, which had been quantified previously, were included in each run. The 160

run was considered valid if the inter-assay coefficient of variation (CV) for the control sample 161

was less than 30%. Each assay was performed in duplicate in the same run. The DNA 162

extraction was considered to be valid if the difference between the cycle threshold (Ct) of the 163

internal control HDV and the average Ct (calculated as the average of the HDV Ct values for 164

all samples in the series) was less than the standard deviation. The reactions, data acquisition 165

and analyses were performed using the ABI PRISM 7000 sequence detection system (Applied 166

Biosystems). Standard curves made from known concentrations of DNA for each set of 167

primers and probes were used for quantification. 168

169

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

8

(iii) Preparation of phylogroup-specific PCR standards. Quantification controls 170

were obtained using one representative strain for each phylogroup [K-12 (A phylogroup), 171

IAI1 (B1 phylogroup), ED1a (B2 phylogroup) and UMN026 (D phylogroup)]. The Genomic 172

DNA Buffer Set (Qiagen) and QIAGEN genomic tips 100 G (Qiagen) were used to extract 173

DNA from bacterial cultures according to the manufacturer’s recommendations. Bacterial 174

counts were obtained by plating, the concentration of purified DNA was determined with a 175

spectrophotometer and the corresponding copy number was calculated according to genome 176

length. The two methods of quantification gave similar results (data not shown). Serial 10-177

fold dilution series for 8-12 to 0.8- 1.2 x 107 target genomes were applied for real-time PCR. 178

179

(iv) E. coli real-time PCR quantification assay. Quantitative PCR for E. coli 16S 180

rDNA was performed as previously described (33) using 20 µl of DNA which was purified 181

from stool samples. 182

183

Control strains. The following bacteria were used to evaluate the specificity of PCR primers 184

and probes: E. coli sensu stricto ED1a, CFT073, 536, S88, IAI1, 55989, IAI39, ECOR36, 185

DAECT14, K-12 MG1655, HS, H10407, 042, UMN026; Escherichia clade I (ROAR 185, 186

M863), clade II (ROAR 19), clade III (ROAR 291, ROAR 438), clade IV (E243, B49), clade 187

V (ROAR 292, ROAR 129) E. fergusonii ATCC35469, E. albertii CIP107988, E. blattae 188

CIP104942, E. vulneris CIP103177, E. hermannii CIP103176, Salmonella enterica 189

Typhimurium, Klebsiella pneumoniae, Proteus mirablis, Pseudomonas aeruginosa 190

ATCC27853, Enterococcus faecalis ATCC33186, Enterococcus faecium, Streptococcus 191

bovis, Bacteroides fragilis, Bacteroides thetaiotamicron, Bacteroides vulgatus, 192

Fusobacterium nucleatum, Prevotella melaninogenica, Clostridium difficile, Clostridium 193

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

9

perfringens, Bifidobacterium adolescentis, Bifidobacterium breve, Eubacterium exiguum, 194

Eubacterium lentum, Lactobacillus casei and Lactobacillus cateniformis. 195

196

Statistical analyses. A factorial analysis of correspondence (FAC) (34) was used to describe 197

associations among the data; a two-way table was analyzed using SPAD.N software (Cisia, 198

Saint Mandé, France). This table had 98 rows (corresponding to the 98 subjects with E. coli 199

identified in their stool samples) and 22 columns (corresponding to the 10 variables). The 10 200

variables were: phylogenetic group (A, B1, B2 or D) of the unique randomly-selected strain, 201

the dominant phylogenetic group, the intermediate phylogenetic group and the minor 202

phylogenetic group (defined below), the presence of high quantitative level of E. coli in stools 203

(more than 108 colony forming units (CFU) per gram), subject gender, subject age (above or 204

below 60 years old) and strain resistance to amoxicillin, ofloxacin or sulfamethoxazole- 205

trimethoprim. We used a binary code for each variable: present =1, absent = 0. The FAC uses 206

a covariance matrix based on χ2 distances. The computation determines a plane defined by 207

two principal axes of the analysis. The first axis, F1, accounts for most of the variance, and 208

the second axis, F2 (orthogonal to F1), accounts for the largest part of the variance not 209

accounted for by F1 (34). The variables of the unique randomly selected strains were used as 210

illustrative variables. All other variables were used to compute the plane. 211

212

RESULTS 213

214

Characteristics of the study population and of randomly selected E. coli commensal 215

strains (one per individual). The study included 46 males and 54 females between the ages 216

of 26 and 86 years (medium 58.0 +/- 11.7). All but two of the subjects carried E. coli. We 217

studied one randomly selected clone per individual; this was considered to be the dominant 218

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

10

clone. Resistance of these 98 E. coli dominant clones to amoxicillin, sulfamethoxazole-219

trimethoprim, amoxicillin-clavulanic acid and ofloxacin was 25%, 14%, 7% and 2%, 220

respectively. No resistance to third generation cephalosporins or amikacin was detected. 221

The classical phylogroup triplex PCR method, used on the dominant clones, showed 222

the prevalence of the different phylogroups to be as follows: A phylogroup (31%), B1 223

phylogroup (13%), B2 phylogroup (33%) and D phylogroup (21%). 224

225

Stool sample quantification of the four main E. coli phylogroups. We developed a real-226

time PCR assay, targeting the genes involved in the classical phylogrouping triplex PCR 227

method (4), to directly quantify the four main E. coli phylogroups (A, B1, B2 and D) from 228

stool samples. In this assay, the phylogroups B1 (chuA negative, yjaA negative and TspE4.C2 229

positive), B2 (chuA positive, yjaA positive and TspE4.C2 variable) and D (chuA positive, 230

yjaA negative and TspE4.C2 variable) were quantified by real-time PCR using probes 231

specifically targeting B1 TspE4.C2, B2 chuA and D chuA sequences, respectively (Table 2). 232

The proportion of each phylogroup was calculated as a percentage of the total E. coli count 233

obtained by quantitative PCR for 16S rDNA (33). The proportion of phylogroup A, which 234

encompasses A0 subgroup (chuA, yjaA and TspE4.C2 negative) and A1 subgroup (chuA 235

negative, yjaA positive and TspE4.C2 negative) (Table 2) was calculated by subtracting the 236

proportions of phylogroups B1, B2 and D from the total E. coli count. The proportion of A1 237

subgroup was confirmed using an yjaA specific probe (pyjaAA1/B2), which quantifies the A1 238

subgroup and the B2 phylogroup using the following formula: A1 = yjaApositive – B2. This 239

pyjaAA1/B2 probe was used to estimate the lower limit of detection of the proportion of 240

phylogroup A. The comparison between the proportion of phylogroup A (calculated by 241

subtraction) and the proportion of A1 subgroup (quantified by the pyjaAA1/B2 probe) allowed 242

us to estimate the threshold of phylogroup A detection by subtraction. Concordance was 243

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

11

observed when the proportion of phylogroup A was at least 15% of the total E. coli count 244

(data not shown). With this approach and using four probes (16S rDNA, pTspE4.C2B1, pchuAB2 245

and pchuAD), we estimated that phylogroup A was actually present at over 15%: and that real-246

time PCR variations did not make a significant contribution to results. This threshold value 247

was further supported by comparison to PCR findings for several clones of plated fecal 248

samples (see below). 249

250

Specificity, sensitivity and reproducibility of the assay. 251

(i) Assay specificity. Specificity of the E. coli primers and probes was confirmed 252

using an extensive set of control strains; this set (mainly gastro-intestinal tract bacterial of 253

non-Escherichia species; see ‘materials and methods’) produced negative PCR results. The 254

specificity of the 16S rDNA assay (33), as well as that of the phylogroup probes and primers, 255

was tested using DNA from various phylogenetic groups of E. coli sensu stricto and from 256

closely related bacterial species of the genus Escherichia [including the recently described 257

Escherichia clades (30, 35)]. The 16S rDNA probe detects the phylogroups of E. coli sensu 258

stricto, E. albertii, E. fergusonii, and the Escherichia clades, but not E. blattae, E. hermanii 259

and E. vulneris, in accordance with their taxonomic status (36, 37). 260

Each phylogroup-specific probe gave positive PCR results for the corresponding target 261

bacteria and negative PCR results for non-target microorganisms, with a small number of 262

exceptions. The Escherichia clade III and IV strains, and some strains of the clade I (ROAR 263

185), were recognized by the chuA B2 probe. The strains of the recently reported phylogroup 264

F, defined by MLST (38), are considered as D1 subgroup by the triplex PCR method (Table 2) 265

but were split into B2 and D phylogroups by the real-time PCR method. This split 266

corresponds to the different branches of the phylogroup F strains on the MLST tree (39). We 267

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

12

did not detect a signal above the threshold for the non-template controls in any of the assays 268

(Ct> 40). 269

Spiking experiments were carried out to test the ability of our assay to pick out a 270

specific bacterial DNA from a complex DNA background. Defined amounts of DNA from 271

each of the studied phylogenetic groups B1, B2 and D were quantified by real-time PCR and 272

spiked with DNA belonging to other phylogenetic group strains. No significant difference 273

was observed, indicating high specificity of the probes used and no evidence of cross-274

reactivity. 275

276

(ii) Assay sensitivity, linearity and efficiency. The detection limits of the real-time 277

PCR assays were examined using 10-fold serial dilutions of target DNAs extracted from pure 278

culture of E. coli A, B1, B2 and D. The sensitivity of the probe specific for pyjaAA1/B2, 279

pchuAB2, pchuAD and pTspE4.C2B1 was below 10 CFU per PCR, which corresponded to 280

approximately 9 x 104 CFU/ g feces. The cycle number at which product fluorescence 281

exceeded a defined threshold varied linearly over a range of concentrations from 106 to 10 282

bacteria per PCR mixture (r2 = 0.99) (data not shown). The linear range of all assays was 283

therefore 9 x 109 to 9 x 104 bacteria per g of feces. The efficiencies of the PCR reactions were 284

around 95% as the slopes of the standard curves were close to the optimal theoretical values 285

(see Table S2). 286

287

(iii) Assay reproducibility. Based on the Ct values of fifteen replicates, the intra-288

assay reproducibility was found to be high (CV of Ct values <1%, corresponding to CV of 289

copies < 20%; Table S2). 290

291

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

13

Comparison with PCR typing of individual clones. To test the robustness of our real-time 292

PCR results we compared our phylogroup quantification to the triplex PCR typing of 20-30 293

clones from 20 subjects. From the number of randomly selected colonies, the probability of 294

detecting a minor clone (defined as a clone constituting up to 10% of the population of E. 295

coli) is estimated to be > 90% (20). We found a correlation between our quantitative PCR 296

assay and the phylogroup determination of individual clones. In all cases, the phylogroups 297

detected by the classical PCR typing of individual clones were also detected by the 298

quantitative PCR assay; we found similar quantitative trends but some differences in the 299

proportions of the phylogroups. Interestingly, the real-time PCR assay detected minor clones, 300

not found by plating, in 40% of the cases (Table S3). 301

302

Quantitative determination of E. coli phylogenetic group composition. Preliminary global 303

quantification of E. coli in the stools of the subjects was carried out using real-time PCR with 304

the 16S rDNA E. coli probe and primers (33). Negative real-time PCR results were found for 305

the two subjects for whom E. coli also failed to be detected by plating. Excluding these two 306

subjects, we found an average of 7.84 +/- 0.54 log CFU of E. coli per gram of stool. The 307

respective proportions of the four phylogenetic groups A, B1, B2 and D was determined for 308

the 98 subjects where E. coli was detected. To verify that the use of transportation swabs sent 309

by regular mail did not bias E. coli quantification, we compared real-time PCR results 310

obtained from eight fresh control fecal samples and from their corresponding swabs which 311

were collected using the same procedure as in the ‘materials and methods’: no significant 312

difference was found (Table S4) in accordance with recent data on storage conditions of 313

samples (40). However, if possible, analysis of fresh feces should be preferred. 314

The phylogenetic groups A, B1, B2 and D were detected in 74%, 36%, 70% and 32% 315

of the subjects, respectively; the groups A and B2 were more frequently detected than B1 and 316

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

14

D. However, the proportion of the four phylogroups varied widely among the subjects. The 317

within-subject prevalence of each phylogenetic group varied from 0 to 100%. A single, two, 318

three or four phylogenetic groups were found in 21%, 48%, 21% and 8% of the subjects, 319

respectively. When only a single group was detected, this was most commonly phylogroup 320

B2 (62%); phylogroup A was found in 19% of subjects and the phylogroups B1 and D in 321

9.5% if subjects. When defining the proportions of the phylogroups according to three 322

categories as suggested by Schlager et al. (20) (dominant phylogenetic group: > 50% of the 323

population of E. coli; intermediate phylogenetic group: 10 to 50% of the population of E. coli 324

and minor phylogenetic group < 10% of the E. coli population), the proportions of these three 325

categories varied significantly according to the four phylogenetic groups (Table 3). 326

Phylogroups B2 and A were equally prevalent as the dominant group, and in a higher 327

proportions of samples than phylogroups B1 or D. Phylogroup A was the most common 328

intermediate phylogenetic group and phylogroup D was the most common minor 329

phylogenetic group. Minor phylogroups (in proportions less than 10% of the total E. coli 330

population) were detected in 40% of the samples. 331

When the 12 combinations (three categories and four phylogenetic groups) were 332

considered, the proportions of the categories varied according to the phylogenetic group 333

considered. This complex variation is described by a FAC (Fig. 1). On the F1/F2 plane of the 334

FAC, which accounted for 32.7% of the total variance, the dominant phylogenetic group B2 335

was distinguished by the positive values of the first factor whereas the dominant phylogenetic 336

groups B1, A and D were distinguished by the negative values of this factor. The intermediate 337

phylogenetic groups B1, B2 and D were projected on the negative values of the F1 axis and 338

thus appeared to be associated with the dominant phylogenetic groups A and B1. The 339

intermediate phylogenetic group A, and the minor phylogenetic groups A, B1 and D, were 340

projected on the positive values of the F1 axis and thus are associated to the dominant 341

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

15

phylogenetic group B2. This analysis allows us to distinguish between two types of structures 342

within the E. coli intestinal commensal populations, according to the major dominant group 343

B2 or A. The dominant phylogenetic group B2 was associated with the absence of the other 344

groups whereas the dominant phylogenetic group A was associated with variable proportions 345

of other groups. The second axis, F2, characterized the dominant phylogenetic group D by its 346

negative value. No link with the absolute quantity of E. coli was observed as the variable > 347

108 CFU was projected near the origin of the axes and thus was not related to phylogenetic 348

repartition. 349

350

E. coli phylogenetic group composition: relevance of subject characteristics and the 351

phylogroup and antibiotic resistance of a randomly selected clone. There is no significant 352

difference of the phylogenetic group composition according to age or gender of the subjects: 353

these variables were projected near the origin of the axes on the FAC (Fig. 1). 354

In 78% of cases, the dominant phylogenetic group determined by the real-time PCR 355

was identical to the one determined by the Clermont method (4) on one randomly selected 356

clone. In 22 % of the cases, the phylogenetic group of the randomly selected clone was 357

present in the sample but as an intermediate or a minor phylogenetic group. There was a 358

discrepancy in six cases which corresponded to the selected clone being identified as 359

phylogroup D whereas the quantitative PCR identified it to belong to phylogroup B2. We 360

have shown that in all six of these cases, the randomly selected clone belonged to phylogroup 361

F (13): this is supported by the fact that the pchuAB2 probe that we designed is positive for 362

some F group strains (Table S3). The good correlation between the quantitative PCR assay 363

and classical phylogroup determination (based on one randomly selected clone) is shown in 364

the Table 3, where the proportions of the dominant phylogenetic groups (A: 38%, B1: 13%, 365

B2: 37%, D: 11%) are close to the proportions detected in the randomly selected clones (A: 366

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

16

31%, B1: 13%, B2: 33%, D: 21%), as well as on the FAC (Fig. 1). The phylogenetic group of 367

the randomly selected clone (considered as an illustrative variable) is projected close to the 368

dominant phylogenetic group on the plane, with the exception of the D phylogroup variable 369

on the negative values of F2 which is attracted by the dominant phylogroup B2 due to the six 370

cases of discrepancy. Resistance of the randomly selected clones to amoxicillin, 371

trimethoprim/sulfamethoxazole or ofloxacin, were also considered as illustrative variables. 372

These were associated with the dominant A phylogroup for negative values of F1 and 373

distantly related to the dominant B2 phylogroup on the FAC (Fig. 1). 374

375

DISCUSSION 376

377

We have developed an original quantitative PCR assay, based on the classical 378

phylotyping method of Clermont and colleagues (4), to determine relative proportions of the 379

four main E. coli phylogroups: A, B1, B2 and D. On average, 7.84 +/- 0.54 log CFU of E. coli 380

were found per gram of feces, in accordance with previous studies (33, 41, 42). When 381

assessing the population structure of intestinal commensal bacteria in humans, it is necessary 382

to study subjects whose microbiota is not perturbed by external factors such as antibiotic 383

consumption, hospitalization or intestinal disease. We were able to study real commensal E. 384

coli populations as our 100 healthy subjects conformed strict inclusion criteria; we excluded 385

any potential subject with a physiopathological condition which could disturb the intestinal 386

microbiota equilibrium. The level of acquired resistance to antibiotics in our study population 387

was much lower than in that reported in community acquired E. coli urinary tract infections in 388

France (ONERBA 2008, AFORCOPI-BIO network 2007, www.onerba.org): amoxicillin 25% 389

versus 44%, sulfametoxazol-trimethoprim 14% versus 20%, amoxicillin-clavulanic acid 7% 390

versus 24% and ofloxacin 2% versus 11%. This supports our belief that the study populations 391

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

17

are commensal, and in the same range of resistance than a recent study in Paris on fecal 392

strains (43) . 393

The specificity of our assay is satisfactory although it has some limitations. Some 394

strains of the recently reported Escherichia clades (30, 35) gave a signal with the pchuAB2 395

probe which can falsely increase the proportion of phylogroup B2. However, these clade 396

strains being yjaA negative at the opposite of the E. coli B2 strains, a strong B2 signal with a 397

weak or negative yjaA signal should evoke the presence of clade I, III or IV strains. The 398

prevalence of these Escherichia clades in human commensal populations is very low (around 399

2-3%) (35) and thus will not significantly affect the results of our assay. We did not detect 400

any Escherichia clade among the 100 randomly selected clones using the new Clermont 401

phylotyping method (13) (Table S3). Some strains classified as phylogroup D by the classical 402

triplex PCR method, and belonging to the recently reported phylogroup F [identified by 403

MLST (38)], appeared to belong to phylogroup B2 due to a positive signal with the pchuAB2 404

probe. Phylogroup F is closely related to the phylogroup B2 (1). This probe cross-reactivity 405

was responsible of all six discrepancies observed between our quantitative PCR assay and the 406

classical triplex PCR phylogrouping of the randomly selected clones (Table S3). However, it 407

may not significantly alter results as phylogroup F strains have been found to represent less 408

than 8 % of the population in a large cohort of fecal strains from Australia and France (13). 409

We have shown that our assay is highly reproducible with a detection threshold of 105 410

CFU / g of feces; this corresponds to 0.1% of the population of E. coli, much lower than the 411

10% threshold obtained by other sampling methods (20, 26, 21, 23, 29). Previous studies 412

picked 20 to 30 isolates from an agar plate per individual and used a semi-quantitative method 413

of assessment (20, 23, 25). We found a good correlation between their data and that obtained 414

using our quantitative method (Table S3). However, real-time PCR may be faster, easier to 415

perform and more suitable for large sample sizes. In 40% of cases (8/20), real-time PCR 416

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

18

detected phylogroups which were not found by plating. Our specific probes were able to 417

detect groups with very low prevalence to a detection threshold. To reach by the alternative 418

semi-quantitative method a similar level of detection of 0.1% as we did in our assay, a large 419

number of isolates would have to be studied from each fecal sample. The classical qualitative 420

method appeared to underestimate the presence of minor groups. This explains the apparent 421

differences in fecal carriage of the E. coli phylogroups (e.g. the prevalence of the B2 422

phylogroup in the studied population is apparently 33% if only major clones are considered, 423

but 70% if all clones, including minor clones, are considered). We acknowledge that, in 424

giving a global estimation of the structure of the commensal strains, our method does not 425

allow us to study the E. coli clones separately. 426

Each person commonly carries a dominant strain of E. coli that constitutes more than 427

half of the total colonies isolated (other strains are present at various levels) (14, 15, 16, 44, 428

18). Depending on the study, between 92% and 100% of human samples contained a 429

dominant clone representing more than 50% of the E. coli population (17, 20, 21, 23). When 430

we compare the results of the quantitative PCR with those based on a qualitative approach in 431

which a single clone was randomly selected per individual, we obtained 78 % correlation if 432

the prevalence of a given phylogenetic group is 50% (Table S3). Our method ensures correct 433

identification of E. coli dominant strain phylogroup. 434

The prevalence of E. coli phylogroups in the stools of humans has been shown to 435

depend on sex, age, year of sampling, country of origin and diet (1, 12, 45). In our population 436

and using our approach, phylogroups A and B2 were detected in 74% and 70%, respectively, 437

whereas the phylogroups B1 and D were detected less often. Phylogroups A and B2 were 438

overrepresented among the dominant phylogroups (Table 3). These results are in accordance 439

with previous data [reviewed in (1)] and confirm the prominent role played now by 440

phylogroup B2, as well as phylogroup A, in commensal microbiota of French subjects. The 441

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

19

B2 phylogroup strains appear opposed to the B1 phylogroup strains that are more animal 442

specific (9). We found no correlations to be associated with gender or age in our samples. 443

High intra-host diversity was found as two, three and four phylogenetic groups were 444

simultaneously detected in 48%, 21% and 8% subjects, respectively. There was a non-445

random, complex level of phylogroup combinations within single subjects. As illustrate in the 446

FAC analysis (Fig. 1), there is antagonism between phylogroups B1 and B2. This could 447

correspond to different ecological niches defined for example by different diets. Among the 448

most prevalent phylogroups (A and B2), there appears to be distinct population structures. A 449

dominant phylogroup B2 was associated with an absence of other groups whereas the 450

presence of the other phylogroups was detected with a dominant phylogroup A (Fig. 1). These 451

data support the notion that extraintestinal virulence may be a by-product of commensalism: 452

numerous “extraintestinal virulence genes” (coding for adhesins, iron capture systems, toxins 453

and protectins) exhibited by phylogroup B2 strains (6, 46) may have evolved primarily to 454

allow the bacteria to be a good intestinal colonizer (24, 47). When phylogroup B2 is 455

dominant, it does not co-occur with the residence of other phylogroup strains, as has been 456

shown epidemiologically (28, 48, 49) and experimentally (50, 51). Alternatively, dominance 457

of phylogroup B2 may reflect the presence of a specific gut environment, or parasitism by 458

viruses or bacteriocins that favor it. As previously reported (52), an association between the 459

non-B2 phylogroup strains, especially phylogroup A, is observed with antibiotic resistance 460

(Fig. 1). This is of particular interest as the intestinal microbiota plays a critical role in the 461

emergence of antibiotic resistance (53, 54, 55). Further work is needed to investigate the 462

characteristics of the hosts corresponding to the two patterns of dominant phylogroup. This 463

would help identify the ecological forces shaping population structure and better understand 464

the impact of exposure to opportunistic E. coli extraintestinal infections and the emergence of 465

antibiotic resistance. 466

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

20

We have developed a rapid and reliable real-time PCR assay that quantifies E. coli 467

phylogroups directly from stool samples and detects, with high frequency, the presence of 468

sub-dominant clones. The presence of minor clones could explain the fluctuation in the 469

composition of E. coli microbiota within single individuals that may be seen over time. They 470

could also constitute reservoirs of virulent and/or resistant strains. Long-term studies of 471

subject cohorts, using our approach, will allow these notions to be tested. 472

473

ACKNOWLEDGMENTS 474

We are grateful to Hervé Jacquier for providing some of the control strains. 475

476

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

21

477

REFERENCES 478

1. Tenaillon O, Skurnik D, Picard B, Denamur E. 2010. The population genetics of 479

commensal Escherichia coli. Nat. Rev. Microbiol. 8:207–217. 480

2. Russo TA, Johnson JR. 2003. Medical and economic impact of extraintestinal infections 481

due to Escherichia coli: focus on an increasingly important endemic problem. Microbes 482

Infect. 5:449–456. 483

3. Herzer PJ, Inouye S, Inouye M, Whittam TS. 1990. Phylogenetic distribution of 484

branched RNA-linked multicopy single-stranded DNA among natural isolates of Escherichia 485

coli. J. Bacteriol. 172:6175–6181. 486

4. Clermont O, Bonacorsi S, Bingen E. 2000. Rapid and simple determination of the 487

Escherichia coli phylogenetic group. Appl. Environ. Microbiol. 66:4555–4558. 488

5. Bingen E, Picard B, Brahimi N, Mathy S, Desjardins P, Elion J, Denamur E. 1998. 489

Phylogenetic analysis of Escherichia coli strains causing neonatal meningitis suggests 490

horizontal gene transfer from a predominant pool of highly virulent B2 group strains. J. 491

Infect. Dis. 177:642–650. 492

6. Picard B, Garcia JS, Gouriou S, Duriez P, Brahimi N, Bingen E, Elion J, Denamur E. 493

1999. The link between phylogeny and virulence in Escherichia coli extraintestinal infection. 494

Infect. Immun. 67:546–553. 495

7. Duriez P, Clermont O, Bonacorsi S, Bingen E, Chaventré A, Elion J, Picard B, 496

Denamur E. 2001. Commensal Escherichia coli isolates are phylogenetically distributed 497

among geographically distinct human populations. Microbiology 147:1671–1676. 498

8. Gordon DM, Cowling A. 2003. The distribution and genetic structure of Escherichia coli 499

in Australian vertebrates: host and geographic effects. Microbiology 149:3575–3586. 500

9. Escobar-Páramo P, Le Menac’h A, Le Gall T, Amorin C, Gouriou S, Picard B, 501

Skurnik D, Denamur E. 2006. Identification of forces shaping the commensal Escherichia 502

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

22

coli genetic structure by comparing animal and human isolates. Environ. Microbiol. 8:1975–503

1984. 504

10. Obata-Yasuoka M, Ba-Thein W, Tsukamoto T, Yoshikawa H, Hayashi H. 2002. 505

Vaginal Escherichia coli share common virulence factor profiles, serotypes and phylogeny 506

with other extraintestinal E. coli. Microbiology 148:2745–2752. 507

11. Zhang L, Foxman B, Marrs C. 2002. Both urinary and rectal Escherichia coli isolates 508

are dominated by strains of phylogenetic group B2. J. Clin. Microbiol. 40:3951–3955. 509

12. Gordon DM, Stern SE, Collignon PJ. 2005. Influence of the age and sex of human hosts 510

on the distribution of Escherichia coli ECOR groups and virulence traits. Microbiology 511

151:15–23. 512

13. Clermont O, Christenson JK, Denamur E, Gordon DM. 2013. The Clermont 513

Escherichia coli phylo-typing method revisited: improvement of specificity and detection of 514

new phylo-groups. Environ. Microbiol. Rep. 5:58-65 515

14. Sears HJ, Brownlee I, Uchiyama JK. 1950. Persistence of individual strains of 516

Escherichia coli in the intestinal tract of man. J. Bacteriol. 59:293–301. 517

15. Sears HJ, Brownlee I. 1952. Further observations on the persistence of individual strains 518

of Escherichia coli in the intestinal tract of man. J. Bacteriol. 63:47–57. 519

16. Sears HJ, Janes H, Saloum R, Brownlee I, Lamoreaux LF. 1956. Persistence of 520

individual strains of Escherichia coli in man and dog under varying conditions. J. Bacteriol. 521

71:370–372. 522

17. Lidin-Janson G, Kaijser B, Lincoln K, Olling S, Wedel H. 1978. The homogeneity of 523

the faecal coliform flora of normal school-girls, characterized by serological and biochemical 524

properties. Med. Microbiol. Immunol. 164:247–253. 525

18. Caugant DA, Levin BR, Selander RK. 1981. Genetic diversity and temporal variation in 526

the E. coli population of a human host. Genetics 98:467–490. 527

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

23

19. Caugant DA, Levin BR, Selander RK. 1984. Distribution of multilocus genotypes of 528

Escherichia coli within and between host families. J. Hyg. 92:377–384. 529

20. Schlager TA, Hendley JO, Bell AL, Whittam TS. 2002. Clonal diversity of Escherichia 530

coli colonizing stools and urinary tracts of young girls. Infect. Immun. 70:1225–1229. 531

21. Lautenbach E, Bilker WB, Tolomeo P, Maslow JN. 2008. Impact of diversity of 532

colonizing strains on strategies for sampling Escherichia coli from fecal specimens. J. Clin. 533

Microbiol. 46:3094–3096. 534

22. Moreno E, Andreu A, Pigrau C, Kuskowski MA, Johnson JR, Prats G. 2008. 535

Relationship between Escherichia coli strains causing acute cystitis in women and the fecal E. 536

coli population of the host. J. Clin. Microbiol. 46:2529–2534. 537

23. Moreno E, Johnson JR, Pérez T, Prats G, Kuskowski MA, Andreu A. 2009. Structure 538

and urovirulence characteristics of the fecal Escherichia coli population among healthy 539

women. Microbes Infect. 11:274–280. 540

24. Nowrouzian FL, Wold AE, Adlerberth I. 2005. Escherichia coli strains belonging to 541

phylogenetic group B2 have superior capacity to persist in the intestinal microflora of infants. 542

J. Infect. Dis. 191:1078–1083. 543

25. Vollmerhausen TL, Ramos NL, Gündogdu A, Robinson W, Brauner A, Katouli M. 544

2011. Population structure and uropathogenic virulence-associated genes of faecal 545

Escherichia coli from healthy young and elderly adults. J. Med. Microbiol. 60:574–581. 546

26. Anderson MA, Whitlock JE, Harwood VJ. 2006. Diversity and distribution of 547

Escherichia coli genotypes and antibiotic resistance phenotypes in feces of humans, cattle, 548

and horses. Appl. Environ. Microbiol. 72:6914–6922. 549

27. Escobar-Páramo P, Grenet K, Le Menac’h A, Rode L, Salgado E, Amorin C, 550

Gouriou S, Picard B, Rahimy MC, Andremont A, Denamur E, Ruimy R. 2004. Large-551

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

24

scale population structure of human commensal Escherichia coli isolates. Appl. Environ. 552

Microbiol. 70:5698–5700. 553

28. Nowrouzian FL, Adlerberth I, Wold AE. 2006. Enhanced persistence in the colonic 554

microbiota of Escherichia coli strains belonging to phylogenetic group B2: role of virulence 555

factors and adherence to colonic cells. Microbes Infect. 8:834–840. 556

29. De Muinck EJ, Øien T, Storrø O, Johnsen R, Stenseth NC, Rønningen KS, Rudi K. 557

2011. Diversity, transmission and persistence of Escherichia coli in a cohort of mothers and 558

their infants. Environmental Microbiology Reports 3:352–359. 559

30. Walk ST, Alm EW, Gordon DM, Ram JL, Toranzos GA, Tiedje JM, Whittam TS. 560

2009. Cryptic lineages of the genus Escherichia. Appl. Environ. Microbiol. 75:6534–6544. 561

31. Le Gal F, Gordien E, Affolabi D, Hanslik T, Alloui C, Dény P, Gault E. 2005. 562

Quantification of hepatitis delta virus RNA in serum by consensus real-time PCR indicates 563

different patterns of virological response to interferon therapy in chronically infected patients. 564

J. Clin. Microbiol. 43:2363–2369. 565

32. Untergasser A, Nijveen H, Rao X, Bisseling T, Geurts R, Leunissen JAM. 2007. 566

Primer3Plus, an enhanced web interface to Primer3. Nucleic Acids Res. 35:W71–74. 567

33. Penders J, Thijs C, Vink C, Stelma FF, Snijders B, Kummeling I, van den Brandt 568

PA, Stobberingh EE. 2006. Factors influencing the composition of the intestinal microbiota 569

in early infancy. Pediatrics 118:511–521. 570

34. Greenacre M. 1992. Correspondence analysis in medical research. Stat Methods Med 571

Res 1:97–117. 572

35. Clermont O, Gordon DM, Brisse S, Walk ST, Denamur E. 2011. Characterization of 573

the cryptic Escherichia lineages: rapid identification and prevalence. Environ. Microbiol. 574

13:2468–2477. 575

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

25

36. Hartl DL. 1992. Population genetics of microbial organisms. Curr. Opin. Genet. Dev. 576

2:937–942. 577

37. Priest FG, Barker M. 2010. Gram-negative bacteria associated with brewery yeasts: 578

reclassification of Obesumbacterium proteus biogroup 2 as Shimwellia pseudoproteus gen. 579

nov., sp. nov., and transfer of Escherichia blattae to Shimwellia blattae comb. nov. Int. J. 580

Syst. Evol. Microbiol. 60:828–833. 581

38. Jaureguy F, Landraud L, Passet V, Diancourt L, Frapy E, Guigon G, Carbonnelle E, 582

Lortholary O, Clermont O, Denamur E, Picard B, Nassif X, Brisse S. 2008. Phylogenetic 583

and genomic diversity of human bacteremic Escherichia coli strains. BMC Genomics 9:560. 584

39. Clermont O, Olier M, Hoede C, Diancourt L, Brisse S, Keroudean M, Glodt J, 585

Picard B, Oswald E, Denamur E. 2011. Animal and human pathogenic Escherichia coli 586

strains share common genetic backgrounds. Infect. Genet. Evol. 11:654–662. 587

40. Lauber CL, Zhou N, Gordon JI, Knight R, Fierer N. 2010. Effect of storage conditions 588

on the assessment of bacterial community structure in soil and human-associated samples. 589

FEMS Microbiol. Lett. 307:80–86. 590

41. Mitsuoka T, Hayakawa K. 1972. The fecal flora of man. I. Communication: the 591

composition of the fecal flora of different age groups. Zentralbl. Bakteriol. Orig. A. 223:333–592

342. 593

42. Furet J-P, Firmesse O, Gourmelon M, Bridonneau C, Tap J, Mondot S, Doré J, 594

Corthier G. 2009. Comparative assessment of human and farm animal faecal microbiota 595

using real-time quantitative PCR. FEMS Microbiol. Ecol. 68:351–362. 596

43. Nicolas-Chanoine M-H, Gruson C, Bialek-Davenet S, Bertrand X, Thomas-Jean F, 597

Bert F, Moyat M, Meiller E, Marcon E, Danchin N, Noussair L, Moreau R, Leflon-598

Guibout V. 2013. 10-Fold increase (2006-11) in the rate of healthy subjects with extended-599

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

26

spectrum β-lactamase-producing Escherichia coli faecal carriage in a Parisian check-up 600

centre. J. Antimicrob. Chemother. 68:562–568. 601

44. Bettelheim KA, Faiers M, Shooter RA. 1972. Serotypes of Escherichia coli in normal 602

stools. Lancet 2:1223–1224. 603

45. Skurnik D, Bonnet D, Bernède-Bauduin C, Michel R, Guette C, Becker J-M, Balaire 604

C, Chau F, Mohler J, Jarlier V, Boutin J-P, Moreau B, Guillemot D, Denamur E, 605

Andremont A, Ruimy R. 2008. Characteristics of human intestinal Escherichia coli with 606

changing environments. Environ. Microbiol. 10:2132–2137. 607

46. Escobar-Páramo P, Clermont O, Blanc-Potard A-B, Bui H, Le Bouguénec C, 608

Denamur E. 2004. A specific genetic background is required for acquisition and expression 609

of virulence factors in Escherichia coli. Mol. Biol. Evol. 21:1085–1094. 610

47. Le Gall T, Clermont O, Gouriou S, Picard B, Nassif X, Denamur E, Tenaillon O. 611

2007. Extraintestinal virulence is a coincidental by-product of commensalism in B2 612

phylogenetic group Escherichia coli strains. Mol. Biol. Evol. 24:2373–2384. 613

48. Ostblom A, Adlerberth I, Wold AE, Nowrouzian FL. 2011. Pathogenicity island 614

markers, virulence determinants malX and usp, and the capacity of Escherichia coli to persist 615

in infants’ commensal microbiotas. Appl. Environ. Microbiol. 77:2303–2308. 616

49. Nowrouzian FL, Oswald E. 2012. Escherichia coli strains with the capacity for long-617

term persistence in the bowel microbiota carry the potentially genotoxic pks island. Microb. 618

Pathog. 53:180–182. 619

50. Adiba S, Nizak C, van Baalen M, Denamur E, Depaulis F. 2010. From grazing 620

resistance to pathogenesis: the coincidental evolution of virulence factors. PLoS ONE 621

5:e11882. 622

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

27

51. Diard M, Garry L, Selva M, Mosser T, Denamur E, Matic I. 2010. Pathogenicity-623

associated islands in extraintestinal pathogenic Escherichia coli are fitness elements involved 624

in intestinal colonization. J. Bacteriol. 192:4885–4893. 625

52. Mammeri H, Eb F, Berkani A, Nordmann P. 2008. Molecular characterization of 626

AmpC-producing Escherichia coli clinical isolates recovered in a French hospital. J. 627

Antimicrob. Chemother. 61:498–503. 628

53. Scanvic-Hameg A, Chachaty E, Rey J, Pousson C, Ozoux ML, Brunel E, Andremont 629

A. 2002. Impact of quinupristin/dalfopristin (RP59500) on the faecal microflora in healthy 630

volunteers. J. Antimicrob. Chemother. 49:135–139. 631

54. Andremont A. 2003. Commensal flora may play key role in spreading antibiotic 632

resistance. ASM News 63:601–607. 633

55. Grenet K, Guillemot D, Jarlier V, Moreau B, Dubourdieu S, Ruimy R, Armand-634

Lefevre L, Bau P, Andremont A. 2004. Antibacterial resistance, Wayampis Amerindians, 635

French Guyana. Emerging Infect. Dis. 10:1150–1153. 636

637 638

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

28

TABLE 1. Primers and phylogroup-specific probes 639

640

Target

organism

Primers

and probe Sequence (5’ to 3’)

Target

gene

Reference

or source

E. coli ECFW

ECRV

16S rDNA

CATGCCGCGTGTATGAAGAA

CGGGTAACGTCAATGAGCAAA

FAM-

TATTAACTTTACTCCCTTCCTCCCCGCT

GAA-TAMRA

16S rRNA (33)

E. coli

phylogroup B1

TspFW

TspRV

pTspE4.C2B1

CAACGCTACCTTGGCGTTATC

CCGCTCTCCAGGCAACATC

FAM-ATCAGCGAGGCCGCGCGA-

TAMRA

TspE4.C2 This study

E. coli

phylogroup B2

chuAFW

chuARV

pchuAB2

ATTAACCCGGATACCGTTACCA

CTGCTGCTGATATGTGTTGAGC

FAM-CGTACCGACCCAACCA-MGB

chuA This study

E. coli

phylogroup D

chuAFW

chuARV

pchuAD

ATTAACCCGGATACCGTTACCA

CTGCTGCTGATATGTGTTGAGC

FAM-CGTACCAACCCACCCAA-MGB

chuA This study

E. coli

phylogroups A1

and B2

yjaAFW

yjaARV

pyjaAA1/ B2

CGCCTGTTAATCGCCAATTT

AAAAGAATGCCAGGTTGAACG

FAM-

AAGTTCTGCAAGATCTTGTTCTGCAAC

TCCA-TAMRA

yjaA This study

641

642

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

29

TABLE 2. E. coli phylogenetic group and subgroup determination using the results of 643

PCR amplification of chuA gene, yjaA gene and DNA fragment TspE4.C2 [as in (4)] 644

645

Phylogenetic group

and sub-group of E. coli

chuA yjaA TspE4.C2

A0 - -a -

A1 - + b -

B1 - - +

B22 + + -

B23 + + +

D1 + - -

D2 + - +

aAbsence of the sequence. 646 bPresence of the sequence. 647

648

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

30

649

TABLE 3. Distribution of phylogenetic groups according to their proportions determined by 650

quantitative PCR of fecal samples from 100 healthy volunteers 651

652

DPGa IPGb MPGc

A 38 27 35

B1 13 10 77

B2 37 13 50

D 11 5 84

aDominant phylogenetic group (> 50%); bIntermediate phylogenetic group (10-50%); c Minor 653

phylogenetic group (<10%) as in (20). 654

655

656

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

31

FIGURE LEGEND 657

658

FIG 1. Factorial analysis of correspondence (FAC) for the 98 subjects with E. coli found in 659

their feces. Projections on the plane F1/F2 of the four phylogenetic groups (A, B1, B2, D) of 660

the unique randomly selected strains, of the dominant phylogenetic group (DA, DB1, DB2, 661

DD), of the intermediate phylogenetic group (IA, IB1, IB2, ID) and of the minor phylogenetic 662

group (MA, MB1, MB2, MD) defined for each of the four phylogenetic groups. The 663

dominant, intermediate and minor groups were defined from data from the quantitative PCR 664

assay. The abundance of E. coli in stools (CFU>108), the gender of the subjects (male or 665

female), the age of the subjects (older or younger than 60 years) and the resistance of the 666

unique randomly selected strains to amoxicillin (AMX-R), ofloxacin (OFL-R) and 667

sulfamethoxazole-trimethoprim (SXT-R) were also projected on the plane. 668

669

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from

on March 15, 2018 by guest

http://aem.asm

.org/D

ownloaded from